Peak Catch and Recent Trend

This function finds, for the total catch and for the main species, countries or fishing areas, the year and size of the peak catch, the catch in the last year as a percentage of the peak, and the trend over the last recent_years years: the average annual growth rate, and a Mann-Kendall trend test (Kendall's tau between year and catch).

Usage

peak_trend(
  area_data,
  by = c("species", "country", "area"),
  top_n = 10,
  recent_years = 10,
  plot = TRUE,
  table = TRUE,
  lang = NULL,
  pptx = getOption("faofishingreport.pptx", FALSE),
  plot_show = fao_setting("plot_show"),
  label = NULL
)

Arguments

area_data: FAO catch data: fishing_area_data()$area_data, or any subset of load_fao_data() (e.g. one country, or several fishing areas).
by: What to compare: "species" (default), "country" or "area".
top_n: The number of species, countries or areas with the largest total catch to include. Default is 10; use Inf for all.
recent_years: The number of most recent years used for the trend. Default is 10.
plot: A logical value. If TRUE (default), charts are saved in figures/png/summary/.
table: A logical value. If TRUE (default), tables are saved in tables/.
lang: The language of the chart text and names: "en" (English), "zh" (Chinese) or "both" (Chinese and English). Default is NULL, which uses the language set with fao_settings() ("en" unless changed).
pptx: A logical value. If TRUE, charts are also saved as editable PowerPoint files in figures/pptx/summary/. Default is FALSE, which can be changed for all functions with fao_settings(pptx = TRUE).
plot_show: A logical value. If TRUE, each chart is also drawn in the plot window (e.g. the Plots pane of RStudio) as it is saved, to look at it at once. Default is FALSE, which can be changed for all functions with fao_settings(plot_show = TRUE). The charts are saved as files (plot = TRUE) either way.
label: The name used in the output file names. Default is the fishing area (or the country) of the data.

Value

Invisibly, a data frame with one row per series (the first row is the total).

Details

Catch series are autocorrelated, which makes the usual Mann-Kendall test find too many trends, so the variance of the test is corrected with the effective sample size of an AR(1) process (Yue and Wang 2004), from the lag-1 autocorrelation of the detrended series corrected for its small-sample bias (Kendall 1954; lag1_autocorrelation); p_value_iid is the p value of the usual test, for comparison. Over ten years the autocorrelation is poorly known and the correction is only partial: strongly autocorrelated series still show trends too often. trend tests each series on its own (p_value < 0.05): with many series, some trends are found by chance, and q_value gives the p values adjusted for the number of series in the result, the total excepted (Benjamini and Hochberg 1995). Read trends with p_value < 0.05 as exploratory, and give as significant the trends with q_value < 0.05, as fao_report() does for the main species and countries.

References

Benjamini, Y. and Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B 57: 289-300. doi:10.1111/j.2517-6161.1995.tb02031.x

Kendall, M.G. (1954). Note on bias in the estimation of autocorrelation. Biometrika 41: 403-404. doi:10.1093/biomet/41.3-4.403

Yue, S. and Wang, C.Y. (2004). The Mann-Kendall test modified by effective sample size to detect trend in serially correlated hydrological series. Water Resources Management 18: 201-218. doi:10.1023/B:WARM.0000043140.61082.60

Examples

if (FALSE) { # \dontrun{
FAO27 <- fishing_area_data("Atlantic, Northeast", plot = FALSE)
peak_trend(FAO27$area_data, by = "species", lang = "zh")
peak_trend(FAO27$area_data, by = "country", lang = "zh")
} # }